Oh, I’m not gonna disagree with you there - the current approach is certainly wrought with obvious problems - similar devices, especially simpler ones, are likely to have the same fingerprint. It was a fingerprinting mismatch that led me to learn about the (not really a) COMMAND_CLASS_MARK. Later, I see that @duncan mentioned 0xEF in the forums, but the docs mention only the single-digit separator. I’m still unsure if either method of separating supported and controlled CCs could be valid for Z-Wave devices with both supported and controlled CCs, or if it’s changed, or if the docs are wrong and the text is a holdover from adapting Zigbee fingerprinting to Z-Wave.
I understand why they did it - the NIF is guaranteed at pairing, and they try to match up device and hander from what they know they’ll have available. But they’re aware of the limitations of this approach as well.
As duncan has said in another thread,
ST appears to be in tune with this, though I’m not sure when we might see this, or if it’s already present and just undocumented as of yet.
Congrats on your device type release and pending publishing!